EXPLORING THE POTENTIAL OF MULTI-TEMPORAL TERRAIN ANALYSIS FOR GEOLOGIC MAPPING AND GEOHAZARD STUDIES IN THE EASTERN UNITED STATES

One of many challenges of geologic mapping and geohazard studies in the Eastern United States is the region’s dense and pervasive vegetative canopy, which obscures terrain features and patterns. The advent of lidar-based high-resolution digital elevation models (DEMs) in the early 2000s largely overcame this challenge – improving the detail and accuracy of terrain mapping, while also paving the way for terrain change to become an important element of geo-studies. Particularly for landslide investigations, multi-temporal terrain analysis greatly improves understandings of slope processes. Unfortunately, multi-date lidar coverage is limited to small, typically urban, areas. The need to “go back in time” for terrain data has been tentatively filled by historical DEMs created through structure-from-motion (SfM) analysis of historical aerial imagery. Such imagery extends back into the 1930s, and at scales of 1:15,000 to 1:25,000 can be used to develop DEMs with resolution comparable to lidar. In the past decade, unmanned aerial systems (UAS) imaging has provided an approachable means of acquiring high-resolution DEMs – again using SfM. However, vegetation still poses a significant challenge for multi-temporal terrain analysis using historical or UAS imagery. Recent technological advances enable the collection of lidar data by UAS, paving the way for new and repeat collection of high-resolution terrain data for small areas. The question remains: ‘How comparable are SfM-, airborne lidar-, and UAS-borne lidar DEMs in multitemporal terrain analysis?’

This presentation summarizes take-aways of multi-temporal terrain mapping from several studies: in Sleeping Bear Dunes, MI multi-date historical DEM analysis improved understanding of the progression of the large landslide at Sleeping Bear Point; in Hendersonville, VA UAS imagery was used to map a large landslide; in Pittsburg, PA historical DEMs created from archival aerial imagery improved visual interpretation of landslide features detected by lidar-lidar terrain differencing; in the Delaware Water Gap, PA a lidar-to-lidar comparison demonstrated the vertical and horizontal accuracy required for landslide and rockfall mapping; finally, in Vicksburg, MS, UAS-borne lidar was compared to airborne lidar to investigate landslide and gullying processes.